Environ. Sci. Technol. 1985, 19, 193-198
B.; Mabey, W.; Holt, B.; Gould, C. “Water Related Environmental Fate of 129 Priority Pollutants”;US.Environmental Protection Agency: Washington, DC, 1979;EPA 440/4-79-029a,Vol. 1 and 2. (21) Ahlberg, M.; Bergheim, L.; Nordberg, G.; Persson, S.; Rudling, L.; Steen, B. Environ. Health Perspect. 1983, 47, 85-102.
(22) Anderson, 0. Environ. Health Perspect. 1983,47,239-253.
Received for review February 17, 1984. Revised manuscript received September 24, 1984. Accepted October I, 1984. This work was performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under Contract W-7405-Eng-48.
NOTES Influence of pH and Ionic Strength on the Aqueous-Nonaqueous Distribution of Chlovnated Phenols John C. Westall” Department of Chemistry, Oregon State University, Corvallis, Oregon 9733 1
Chrlstlan Leuenberger and Reni P. Schwarrenbach Swiss Federal Institute for Water Resources and Water Pollution Control (EAWAG), CH-8600 Duebendorf, Switzerland
H The distribution ratio of hydrophobic ionizable organic
compounds (HIOC’s) between aqueous and nonaqueous phases is shown to depend on the pH and ionic strength of the aqueous phase. Four models are presented to describe the association of the HIOC with the nonaqueous phase: (i) transfer of the neutral organic species from the bulk of the aqueous phase to the bulk of the nonaqueous phase; (ii) transfer of the ionic organic species with inorganic counterions to the bulk of the nonaqueous phase; (iii) transfer of the ionic organic species to the aqueous-nonaqueous interface with inorganic counterions in the aqueous phase; (iv) association of the organic species with specific functional groups of the nonaqueous phase. The distribution of 2,3,4,5-tetrachlorophenoland pentachlorophenol in the two-phase system composed of KOH, KC1, H20, octanol, and the chlorophenol was determined as a function of ionic strength and interpreted quantitatively in terms of models i and ii. For aqueous phases with high pH values and ionic strengths, the dominant species of the chlorophenols in the octanol phase were the chlorophenolate ions in association with K+ counterions. Introduction The distribution ratios of nonpolar organic compounds between water and natural sorbents have been estimated with satisfactory accuracy from their relationships to the organic carbon content of the sorbent and the octanolwater partition constant of the solute. These correlations have been of great value in the estimation of the degree of sorption of a wide variety of compounds on a wide variety of sorbents. However, it has been noted in some of the correlation studies that polar and especially ionizable organic compounds do not fit to the same pattern as the nonpolar compounds. In a recent study on the sorption of chlorinated phenols by sediments and aquifer materials, Schellenberget al. (1) demonstrated that sorption of not only the neutral phenol but also the anionc phenolate occur. In certain cases the phenolate species contributes significantly to the overall distribution ratio of the compounds. Kaiser and Vald0013-936X/85/0919-0193$01.50/0
Table I. Properties of Some Hydrophobic Ionizable Organic Compounds compound 2-chlorophenol 2,4-dichlorophenol 2,4,6-trichlorophenol pentachlorophenol 2-nitrophenol 4-nitrophenol 2,4-dinitrophenol
2,4-dimethylphenol 3-methyl-4-chlorophenol 2-methyl-4,6-dinitrophenol
log K,”
log K,”
8.52 7.85 5.99 4.74 7.21 7.15 4.09 10.60
2.17 2.75 3.38 5.01 1.76 1.91 1.53 2.50 2.95 2.85
4.35
nValues from Callahan et al. (3). ~~~
~
~~
manis (2) have reported the dependence on pH of the distribution ratio of pentachlorophenolbetween water and octanol, but no systematic study of the effect of ionic strength was made. Experimental results of Schellenberg et al. (I) indicate that phenolate sorption is strongly influenced by the organic carbon content of the sorbent and by the ionic composition of the aqueous phase. The authors conclude that a much more detailed understanding of the interactions that govern the distribution of such hydrophobic anions between aqueous and nonaqueous phases is necessary to arrive at a quantitative description of the sorption of chlorinated phenols and other hydrophobic ionizable organic compounds (HIOC’s) by natural sorbents. It is the purpose of this paper to examine the theoretical framework on which the distribution of HIOC’s between aqueous and nonaqueous phases is to be viewed and to test the suitability of this theory in the octanolwater system with two HIOC’s of environmental concern, pentachlorophenol and 2,3,4,5-tetrachlorophenol. Although only two HIOC’s are considered in this paper, several of the 129 priority pollutants discussed by Callahan et al. (3) belong to this class of compounds, as is shown in Table I. The KO,referred to in the table is the partition constant for the neutral species. As is seen for the chlorophenols, an increase in the number of chloro substituents
0 1985 American Chemical Society
Environ. Sci. Technol., Vol. 19, No. 2, 1985
193
Table 11. Association of Organic Compound with Nonaqueous Phasea
I. solvation of neutral organic compound in bulk nonaqueous phase: A H = AH 11. solvation of ion pair or free ions in bulk nonaqueous phase: A- + K+ = + F) 111. adsorption of organic ion from aqueous phase onto lipophilic surface with counterion in electric double layer: E + A - + K+==+ K* IV. ligand exchange of organic molecule with surface hydroxyl group a t inorganic oxide surface: EXOH + A H = EXA t H,O a Symbols: AH represents neutral pentachlorophenol; A- represents pentachlorophenolate ion; XOH represents an oxide surface hydroxyl group, for example, :SiOH, :AlOH, and :FeOH; the overbar indicates species in the nonaqueous phase ; E represents surface (nonaqueous phase).
Table 111. Possible Species in the System: H20, KOH, HCl, AH, and Octanol nonaqueous phase ion pairs ions
aqueous phase
AH -
HzO H+ OHK+
(=r
AK
KCI -
c1-
KOH -
AH A-
H+ AK+ c1-
OH-
HCl
For many nonpolar organic compounds only one species of the compound exists, and the values of Kp and D are identical. The concentrations of the neutral and ionized species in the aqueous phase are related by the acidity constant, hydrogen ion concentration, and activity coefficients:
K, =
[A-I [H+I YA-YH+
(3)
[AHITAH
leads to an increase in the acidity and an increase in the hydrophobicity. Thus, the phenols with more chloro substituents are more likely to be ionized at environmental pH values and are intrinsically more hydrophobic. However, the same effect is not seen for the nitrophenols, since the nitro substituent, which causes the increase in acidity of the phenol, causes much less hydrophobicity than the chloro substituents. These examples point out that the class of HIOC’s includes many compounds of environmental concern and that there is reason to include the ionic species in studies of the transport and the fate of these compounds.
Theory Several mechanisms can be considered for the association of the HIOC with the nonaqueous phase. The relative importance of each mechanism depends primarily on the K , and KO, of the organic molecule, the pH and ionic strength of the aqueous phase, and the nature of the nonaqueous phase. The experimental work presented here indicates that free ions or ion pairs can be the predominante forms of the organic compound in the nonaqueous phase at pH values favoring the ionized form in the aqueous phase. Four mechanisms by which organic solutes associate with the nonaqueous phase are presented in Table 11. Only mechanisms I and I1 will be considered in interpretation of the experimental data, but mechanisms I11 and IV are presented for comparison. To facilitate discussion of the mechanisms, a specific system is considered: distribution of pentachlorophenol (PCP represents the compound in general, AH represents specifically the neutral species, and A- represents the anionic species) between a nonaqueous phase and an aqueous phase containing water, a salt (KCl), an acid (HCl), and a base (KOH). A distinction must be made between the partition constant, Kp, which is defined for a particular species:
and the distribution ratio, D, which is defined for the total analytical concentration:
D= 194
[AH] + [ E ] [AH] + [A-]
[m]
=-
[PCPI
Environ. Sci. Technol., Vol. 19, No. 2, 1985
Mechanism I of Table I1 involves the distribution of a neutral species between the bulk of the aqueous phase and the bulk of the nonaqueous phase, with no side reactions. This mechanism and the linear free energy relationships described by Leo et al. ( 4 )have been used successfully to interpret energetics of sorption for a wide variety of nonpolar organic compounds (5-8). The partition constant K:J of an organic compound i between water and any natural sorbent j can be calculated from the octanol-water partition constant of the compound Kowiand the fraction organic carbon of the sorbent fo> log KJ:
= log KO:
+ log fa?
+ + log fa?
= a logKowi b
(4) The intermediate quantity KO: is the partition constant for a hypothetical sorbent of 100% organic carbon. For a particular compound and sorbent, the partition constant is the ratio of concentration in the aqueous phase to the nonaqueous phase. For compounds that yield both neutral and ionic species in aqueous solutions, the distribution ratio D can be calculated from the value of K p given from eq 4 and the acidity constant (5)
if it is assumed that only the neutral species is present in the nonaqueous phase and activity corrections are neglected. In many systems it is clear that ionic species do enter the nonaqueous phase, and more complete mechanisms have to be considered. Mechanism I1 of Table I1 is simply the mechanism studied in solvent extraction chemistry, only with the emphasis placed on extraction of the organic compound rather than on extraction of the inorganic ion. This mechanism can be illustrated by considering equilibrium in the system composed of HzO, HC1, KOH, AH, and octanol, i.e., the partition of PCP between an aqueous phase of variable pH and ionic strength and nonaqueous phase of water-saturated octanol. The species present in the aqueous phase are known with certainty (Table 111). In the nonaqueous phase, the species H20 and AH are certain to exist. The experiments reported here give evidence for the presence of organic and inorganic ions or ion pairs in
Table IV. Reactions in System HzO, HCl, KOH, PCP, and Octanol
-
log Ka (298 K, I 0)
reaction
am
(1) K+ + C1- = ii'+ + A(2) K+ + A- = (3) K+ + OH- = ii'+ (4) H+ + A- = 9 + (5) H+ + Cl- = 9 + (6) H+ + OH(7) K+ + C1- = KC1 (8) K+ + A- = AK (9) K+ + OH- = KOH (10)AH = AH (11) HCl = HCI (12) HzO = HzO (131 HoO H+ + OH(i4j AH = H+ + A-
-7.9b -7.1"
=*+m -2.gb 2.6c 5.24d
-14.00 -4.75d
-
= For extrapolation of equilibrium constants to I 0 the Davies equation was used to estimate aqueous phase activity coefficients. Nonaqueous phase activity coefficients were set equal to 1. *Determined from data in Figure 2 (lower curve) with a modified form of FITEQL (IO), a weighted nonlinear least-squares optimization procedure for determination of constants in chemical equilibrium systems. Residuals in K+ concentration in octanol were minimized with respect to constants for reactions 1 and 7, only. Reactions 3 and 9 were found to be less significant under prevailing experimental conditions and, along with reactions 5, 6, and 11, were not included in the model. Determined from data in Figure 1 with a modified form of FITEQL (IO). Weighted residuals in total mass of PCP in the system and total concentrations of PCP in octanol and in water were minimized with respect to constants for reactions 2 and 8 and the constraint of material balance of PCP in the system. dFrom ref 1.
the nonaqueous phase. Since the nonaqueous phase in this system contains -3.2 M HzO, solvation of free ions and their presence in the nonaqueous phase is not implausible, as will be discussed. For the sake of completeness, all of the possible ion pairs and free ions in the organic phase are listed in Table 111, even for species likely to be present at very low concentrations. The set of reactions that completely describes the formation of these species is given in Table IV, along with values for equilibrium constants estimated from this work. All of the reactions listed in Table IV are electroneutral; i.e., they involve no net transfer of charge from the aqueous phase to the nonaqueous phase. If transfer of single ions were considered, an extrathermodynamic assumption would have to be made with regard to how the observed energy of transfer for an electroneutral pair should be divided between the anion and the cation. Mechanism I1 has been approached as a multicomponent chemical equilibrium problem. The speciation of solutes in the octanol-water system containing HC1, KOH, and AH is uniquely and completely determined by the material balance equations for the solutes, the electroneutrality equation for each phase, and the mass action equations for the reactions in Table IV. This equilibrium problem can be solved by the computer program MICROQL (91, and equilibrium constants can be determined from experimental data with the program FITEQL (IO). Mechanism I11 involves the removal of the HIOC to a lipophilic organic surface. For ionized organic compounds, the counterion must no longer by transferred to the organic phase but may remain in the electric double layer in the aqueous phase. In this case the HIOC's behave as ionic surfactants. In natural systems there is certainly a continuous transition between mechanism I1 (transfer to a
three-dimensional organic phase) and mechanism I11 (transfer to a two-dimensionalorganic surface). However, it is instructive to examine laboratory results for a purely mechanism I11 reaction as a limiting case. Cantwell and Puon (11)found the adsorption of diphenylguanadinium (DPGHf) ion on XAD-2 (a polystyrene-divinylbenzene resin) to be describable in terms of a Gouy-Chapman model of the interface with C1- ions as counterions in the electric double layer. For constant surface potential and constant bulk concentration of DPGH+, the concentration of adsorbed DPGH+ was found to vary with the square root of the ionic strength as predicted by the GouyChapman theory. Mechanism IV involves the specific interaction of the ionogenic functional group of the organic molecule with a functional group of a purely inorganic surface. This mechanism is unique with respect to the other three in that it involves, at least conceptually, interaction of specific functional groups, as opposed to an undefined interaction of organic solvents and organic sorbents. Mechanism IV can be interpreted in terms of the surface complexation model as reviewed recently by James and Parks (13), Schindler (14),Westall and Hohl (15) and Stumm et al. (16). According to this model hydroxyl groups on the surface of hydrous oxides (silica, alumina, iron and manganese oxides, clays, etc.) react similarly to hydroxide groups in homogeneous solution-complexing cations and exchanging for anions. Stumm et al. (16) have studied the surface complexation of the aromatic compounds benzoic acid, salicyclic acid, phthalic acid, and catechol on r-Al,O, and found detectable removal of all four from solution. These results suggest that the removal of certain ionic organic compounds on pristine oxide surfaces should not be overlooked. However, the results of Schellenberget al. (1) indicate that organic carbon is still the dominant sorbent of chlorophenolate ions in a sediment and a natural aquifer material. In this report only anionic HIOC's are considered. It is to be expected that the relative importance of the various mechanisms of transfer is different for cationic HIOC's, since most natural sorbents are negatively charged at environmental pH values. Experimental Section The distribution ratios of pentachlorophenol, 2,3,4,5tetrachlorophenol, and 3,4,5-trichlorophenol were determined in an aqueous phase composed of water, KCl(0.2, 0.1, and 0.05 M), KOH (0.01 M), and the nonaqueous phase of octanol saturated with the aqueous solution. The chlorophenols were analytical grade from Merck, and the octanol was spectroscopic grade from Merck, all used without further purification. All water was doubly distilled in quartz. HC1 and KOH solutions were prepared from Merck ampules, and KC1 was analytical grade from Merck. Aqueous solutions with a particular concentration of KOH and KC1 were prepared and preequilibrated with a small volume of octanol. Octanol was preequilibrated with a small volume of the aqueous solution. Then volumes of the preequilibrated aqueous and nonaqueous phases were mixed and spiked with a volume of the chlorophenol dissolved in octanol. The solutions were mixed for 24 h at 20 OC on a shaker, the phases were separated, and samples of the aqueous and nonaqueous phases were withdrawn for analysis from the bulk of the solutions with a pipet. The concentrations of the chlorophenolsin the aqueous and nonaqueous phases were determined by ultraviolet spectrophotometry. Absorbances were measured at 11 wavelengths between 280 and 340 nm, and a nonlinear Environ. Sci. Technol., Vol. 19, No. 2, 1985
195
+ Experiment
+ Experiment
0 Model
0 Model
300
Y
t
200
0
2001
'Or
i with
ji/,
I#
without
0 0
0.10 TKCl
0.20
(M)
Figure 1. Concentration of PCP distributed between octanoi and water. Composition of aqueous phase 0.01 M KOH and KCI. Constants for equilibrium model in Table IV.
least-squares calibration and data reduction technique was used to allow determination of concentrations of both neutral and anionic species. Mass balance between material added and material found agreed to within l%in most cases and was no worse than 5% in any case. Volumes of solutions were chosen such that the amount of the chlorophenol in either phase was always at least one-fifth of the total amount added. The concentration of K+ in the nonaqueous phase was determined by standard addition technique with flame atomic absorption spectrophotometry. Samples were diluted to 2 times their original volume with methanol to decrease viscosity of the solution in order to increase the precision of the analysis. The pH of the aqueous solutions was determined with a glass electrode and a Ag/AgCl/KCl reference electrode fitted with a Wilhelm salt bridge. The concentration of KC1 in the reference electrode and salt bridge was the same as that in the aqueous solution. The glass electrode was calibrated to H+ concentration by addition of standard HCl and KOH solutions to solutions of the same concentration of KC1 as the particular sample of which the pH was to be determined. The acidity constants of the chlorophenols were determined by the method described by Schellenberg et al. (I).
Results and Discussion The concentration of PCP in octanol and water as a function of aqueous KC1 concentration is shown in Figure 1. The nonaqueous phase concentration of PCP is seen to be a strong function of KC1 concentration. The concentration of PCP in the aqueous phase remained approximately constant over the range of KC1 concentration, since the volumes of aqueous and nonaqueous phases and amount of PCP added were selected with that goal in mind. The multicomponent chemical equilibrium model (to be discussed) has been used for interpretation of the experimental data; good agreement between the model and the experimental data is seen. In Figure 2 is shown the concentration of K" in the nonaqueous phase as a function of concentration of KC1 in the aqueous phase. It is seen that PCP has a dramatic 196
Environ. Sci. Technoi., Voi. 19, No. 2, 1985
0.10
0.20
TKC,(M)
Flgure 2. Concentration of K+ in octanol with PCP (at concentrations shown in Figure 1) and without PCP. Composition of aqueous phase: 0.01 M KOH, KCi, and PCP.
effect on the extraction of K+ from the aqueous phase. Through comparison with Figure 1 it is seen that the concentration of K+ in the nonaqueous phase follows very closely the concentration of PCP in the nonaqueous phase. It is obvious from the data that there is transfer of inorganic ions (K+)into the nonaqueous phase along with the organic phenolate ion. Also the transfer of salts KC1 or KOH must occur. Whether these inorganic ions exist in the organic phase as ion pairs or free ions is not resolved. The experimental data and calculations based on Bjerrum's model (17)of ion association in a continuous dielectric medium suggest that both free ions and ion pairs could contribute significantly to the total amount of PCP in the nonaqueous phase. Chemical Equilibrium Model. The chemical equilibrium model used to describe the experimental results is completely defined by the material balance equations for PCP, KOH, and KC1, the electroneutrality equation for both phases, and the mass action laws for the reactions given in Table IV. The reactions in the table are written for the transfer of neutral salts from the aqueous phase to the nonaqueous phase. No experiments were carried out to allow estimation of single ion partition constants. Values of_ the_ constants for the reactions representing transfer of K+, C1-, and KC1 (reactions 1 and 7 in Table IV) were found from the data in the lower curve of Figure 2. Other experiments showed that the concentration of K+ in the nonaqueous phase did not vary strongly with pH for constant KC1 concentration in the aqueous phase. On this basis it was decided to omit reactions 3 and 9 from the equilibrium model, although these reactions probably become significant under other experimental conditions. For a simple salt solution such as KC1 the predominance of free ions vs. the predominance of ion pairs in the nonaqueous phase can be assessed from the dependence of nonaqueous phase salt concentration on aqueous phase salt concentration: a predominance of free ions results in a linear dependence, and a predominance of ion pairs results in a quadratic dependence. The formation of ion pairs is favored a t higher salt concentrations. From the linearquadratic form of the lower curve in Figure 2, both the free ion and the ion pair are expected to be significant species. On account of the limited domain over which experimental
3r
t Experiment
I
+
O Mode‘
CKCl
2-
0.2 M
0
-
CII
0.02 M I -
0.002 M
!I2
-11
-10
-9
-8
-7
log [H+l 0
0.10
0.20
TKCl
Figure 3. Concentration of TeCP distributed between octanol and water. Composition of aqueous phase: 0.01 M KOH and KCi. Equilibrium model deflned in Table IV, except with constants for TeCP (78).
data were obtained, the values found for the constants are subject to considerable covariability. However, it is worth noting that the values obtained for the equilibrium constant for ion association in the organic phase
w+X-=MX Kip
(6)
are in the range predicted by Bjerrum’s theory for ion association in a continuous dielectric medium such as octanol or water-saturated octanol. The constants for the other major reactions, the extraction of K+,A- and were determined from the data in Figure 1 and the previously determined values for K+,C1- and AH. Again, both the associated and dissociated forms of K+, A- appear to be of importance. Experiments similar to those described for distribution of pentachlorophenol were carried out with 2,3,4,5-tetrachlorophenol (TeCP). The aqueous phase concentrations are shown in Figure 3. It is seen that transfer to the nonaqueous phase is favored at higher ionic strengths but that tetrachlorophenol is less favored in the nonaqueous phase than is pentachlorophenol. Thus, the order of hydrophobicity of the neutral phenols is maintained for the ionic phenolates. An equilibrium model defined by reactions similar to those of Table IV but with equilibrium constants for TeCP (18) can be used to interpret the experimental data for the partition experiments with TeCP, as seen in Figure 3. The equilibrium model that was developed for PCP (reactions in Table IV,mass balance, and electroneutrality) has been applied to estimate the concentration distribution ratio of PCP at any ionic strength over a range of pH values. The results of such computations are shown in Figure 4. At pH values less than pH 7 the distribution is dominated by the transfer of the neutral species AH and is independent of salt concentration. (The salting out effect which occurs at higher ionic strengths is not considered in this model.) At pH values greater than pH 10, the predominate PCP species in the nonaqueous phase is the phenolate anion, and the distribution ratio is strongly related to salt concentration. Note that the relationship between log D and log CKClis not linear, reflecting the
m,
Figure 4. Distribution ratio of PCP between octanol and water as a function of pH for several lonlc strengths. Computed for equilibrium among reactions listed in Table IV. Aqueous phase activity coefficients were estimated from the Davies equation. I f phenolate ion were not present in the octanol phase (correspondingto C,, N 0), the observed log D vs. log H+ relationship would follow the broken line.
presence of both free phenolate ions and phenolate ions associated with K+ in the nonaqueous phase.
Summary A first step toward resolving the influence of aqueous phase pH and ionic strength on the aqueous/nonaqueous distribution of hydrophobic ionizable compounds has been taken. Evidence has been presented that both free organic ions as well as organic-inorganic ion pairs are significant species in the nonaqueous phase. Thus, aqueous phase composition must be considered in calculating the distribution of these compounds. Further experiments are in progress to elucidate the effects of transfer to a threedimensional nonaqueous phase or transfer to a two-dimensional (surface) nonaqueous phase. Registry No. 2,3,4,5-Tetrachlorophenol,4901-51-3; pentachlorophenol, 87-86-5; octanol, 111-87-5.
Literature Cited (1) Schellenberg, K.; Leuenberger, C.; Schwarzenbach, R. P. Enuiron. Sei. Technol. 1984, 18, 652-657. (2) Kaiser, K. L. E.; Valdmanis, I. Can. J . Chem. 1982, 60, 2104-2106. (3) Callahan, M. A.; Slimak, M. W.; Gabel, N. W.; May, I. P.; Fowler, C.; Freed, J. R.; Jennings, P.; Durfee, R. L.; Whitmore, F. C.; Maestri, B.; Mabey, W. R.; Holt, B. R.; Gould, C. “Water-Related Environmental Fate of 129 Priority Pollutants”; U. S. Environmental Protection Agency: Washington, DC, 1979; Vol. I and 11, EPA-4401 4-79-029a and EPA-440/4-79-029b. (4) Leo, A.; Hansch, C.; Elkins, D. Chem. Rev. 1971, 71, 525-553. (5) Kenaga, E. E.; Goring C. A. I. ASTM Spec. Tech. Publ. 1980, No. 707.
(6) Schwarzenbach, R. P.; Westall, J. Enuiron. Sei. Technol. 1981,15, 1360-1367. ( 7 ) Karickhoff, S. W. Chemosphere 1981, 10, 833-846. (8) Brown, D. S.; Flagg, E. W. J. Enuiron. Qual. 1981, 10, 382-386. (9) Westall, J. “MICROQL: I. A Chemical Equilibrium Program in BASIC”;EAWAG Duebendorf, Switzerland, 1979. (10) Westall, J. “FITEQL. A Computer Program for Determination of Chemical Equilibrium Constants from Experimental Data”. Department of Chemistry, Oregon State Envlron. Sci. Technol., Vol. 19, No. 2, 1985
197
Environ. Sci. Technol. 1985, 19, 198-199
University, Covallis, OR, 1982, Report 82-01. (11) Cantwell, F.; Puon, S. Anal. Chem. 1979,51,623-632. (12) Horvath, C.; Melander, W.; Molnar, I. Anal. Chem. 1977, 49, 142-154. (13) James, R. 0.;Parks,G. Surf. Colloid Sei., 1982,12,119-216. (14) Schindler, P. W. In “Adsorption of Inorganics a t SolidLiquid Interfaces”; Anderson, M.; Rubin, A., Ed.; Ann Arbor, Science Publishers: Ann Arbor, MI, 1981. (15) Westall, J.; Hohl, H. Adv. Colloid Interface Sei. 1980,12, 265-294.
(16) Stumm, W.; Kummert, R.; Sigg, L. Croat. Chem. Acta 1980, 53, 291-312. (17) Harned, H. S.; Owen, B. B. “The Physical Chemistry of Electrolytic Solutions”; American Chemical Society: Washington, DC, 1958. (18) Westall, J., Oregon State University, Corvallis, OR, unpublished data, 1983.
Received for review January 16,1984. Revised manuscript received May 15, 1984. Accepted September 4, 1984.
CORRESPONDENCE Comment on “FishSediment Concentration Ratios for Organic Compounds” SIR: The examination of fish/sediment concentration ratios by Connor ( I ) provides an instructive comparison of the environmental behavior and fate of chlorinated aromatic hydrocarbons (CAHs) and polynuclear aromatic hydrocarbons (PAHs). However, errors in his eq 3 led him to erroneously conclude that “fish/sediment ratios greater than about 0.01 are higher than could be predicted” and that each of the fish liverlsediment ratios in his Figure 3 “is 3-5 orders of magnitude higher than predicted by assuming an organic content of 1% carbon.” In fact, ratios much higher than 0.01 can be predicted, and the ratios in his Figure 3 are very close to the values predicted by the corrected equation. Thus, fishlsediment ratios predicted by using Connor’s approach are much closer to observed field ratios and therefore even more useful than Connor suggested. Two errors are present in Connor’s eq 3. First, in converting the organic-carbon-normalized partition coefficient (KO,) to the mass-normalized, or sediment, partition coefficient (Ksed), one should multiply KO, by f,,, the fraction of the sediment, by weight, composed of organic carbon; Connor expressed f,, as a percent, thereby introducing an error of 2 orders of magnitude into his calculations. The relationship shown in his Figure 1 thus represents ratios for sediments containing 100% organic carbon as stated in the text, not 1% as stated in the figure legend. Second, because K, and f, are in the denominator of the ratio, the log (f,,) term [or, alternatively, log (70 organic carbon/100)] should be subtracted from the right side of eq 3 and not added. This error led Connor to conclude that sediments with less than 100% organic carbon should have a lower fishlsediment concentration ratio, when in fact they are expected to have a higher ratio. The corrected equation for predicting the fish/sediment concentration ratio from Kenaga and Goring’s (2) relationships is log (BCF/Ksed) = -2.872 0.391 log (KO,) - log (f,,.) (1)
+
where BCF is the fish/water bioconcentration factor and KO, is the octanol/water partition coefficient. For an organic carbon content of 1%(Le., f,, = 0.01), the log (BCF/K,,d) ratio increases from 0.30 to 1.67 as log (KO,)varies from 3.0 to 6.5. If the contaminant concentration in fish liver is 25 times higher than the whole-body 198
Environ. Sci. Technol., Vol. 19, No. 2, 1985
concentration, as Connor notes, then the log (fish liver/ sediment) ratio is expected to be 2.87 and 2.39 for a log (KO,) of 6 and f, of 0.01 and 0.03, respectively. These two predicted ratios are near the top and center, respectively, of the cluster of observed values for CAHs shown in Connor’s Figure 3. If the fish/sediment ratios for polychlorinated biphenyls (PCBs) shown in Connor’s Figure 2 are examined separately as freshwater, marine, and laboratory values, the correlation with flushing time is much less evident. The four freshwater ratios, the four largest ratios in the figure, do not show a strong correlation with flushing time. However, they appear to be within an order of magnitude of 45, the steady-state ratio predicted by eq 1, assuming that log (KO,) = 6.47, as reported for Aroclor 1254 (3),and that the sediments are 1% organic carbon. A similar steady-state ratio of 21 is predicted by using the same assumptions and eq 2, based on the equations of Oliver and Niimi (“low exposure”) ( 4 ) and Karickhoff et al. (51, and converting KO,to Ksed: log (BCF/Ksed) = [0.997 log (KO,) - 0.8691 - [log (KO,) - 0.211 - log (f,) (2)
Variation about the predicted fishlsediment ratio is cert,inly to be expected. Kenaga and Goring (2) reported 95% confidence limits of i1.35 for the log (BCF) predicted from log (KO,) and f1.37 for the log (K,) predicted from log (KO,). Also, the slope of the log (fish/sediment) vs. log (KO,) relationship depends upon which equations are used in predicting the ratio. The slope can be positive (eq 1) ( I ) , zero (5,6),nearly zero (eq 2) ( 4 , 5 ) ,or negative (5, 7). Several factors can contribute to variation about the predicted ratio. For example, for certain chemicals, accumulation through the food chain pathway could make the fishlsediment ratio higher than predicted by this analysis (though perhaps still within the 95% confidence limits). I t is not correct to refer to the laboratory bioconcentration factors from the studies cited by Connor ( 2 , 3 , 6 , 8,9) as “96-h” values. Although many chemicals can reach steady-state concentrations in fish in less than 96 h, the more lipophilic chemicals may take several weeks, as Connor notes, and some of the BCFs measured in the cited studies were accordingly made after several weeks of exposure. For example, steady-state BCFs were estimated by Veith et al. (3) from 32-day exposures. Although one technique for rapid estimation of BCFs involves 5-day exposures followed by a depuration period (IO), this technique is used to estimate steady-state BCFs, not 96-h
0013-936X/85/0919-0198$01.50/0
0 1985 American Chemical Society
